STATE OF THE ART
Human neutrophil antigen and antibody studies: a Taiwanese experience
Chen-Chung Chu, Transfusion Medicine Laboratory, Department of Medical Research, Mackay Memorial Hospital, Taipei, Taiwan
Neutrophil alloantibodies are clinically known to cause neonatal neutropenia and transfusion-related lung injuries. Furthermore, neutrophil autoantibodies are associated with autoimmune neutropenia. This article is a review on neutrophil alloantigen in the Taiwanese population, and case investigations related to allo/auto neutrophil antibodies over a period of 20 years in the immunohaematology reference laboratory of Mackay Memorial Hospital, Taiwan.
In Taiwanese populations, HNA-1a is more frequent than HNA-1b (90% and 50% respectively) and HNA-1c is not seen. Furthermore, the FcgRIIIb-deficient individual homozygous for FCGR3B gene deletion is estimated to be approximately 1 in 180 and is not as rare as previously thought.
Although mothers homozygous for either HNA-1a or HNA-1b are commonly found in Taiwan, the number of patients with neonatal neutropenia caused by HNA-1 alloantibodies is seldom seen. In a search among our previous laboratory requests of neonatal cases for granulocyte antibodies testing over a 10-year period, only four cases showed neutrophil alloantibodies in the mother and baby’s sera.
Lastly, primary autoimmune neutropenia among 51 children was retrospectively investigated. The study revealed that the specificity of the autoantibody was most commonly directed against HNA-1a. In addition, 31 cases were genotyped for HLA-DRB1 and -DQB1 genes. The HLA allele HLA-DQB1*05:03 was found to be associated with primary autoimmune neutropenia in Taiwan.
Neutrophil allo/autoantibodies are involved in alloimmune neonatal neutropenia (ANN) and autoimmune neutropenia (AIN), respectively. Furthermore, neutrophil alloantibodies are believed to be implicated in refractoriness to granulocyte transfusions, febrile transfusion reactions and transfusion-related lung injuries (TRALI) . The human neutrophil alloantigens (HNAs) belong to five systems (HNA-1–5), each corresponding to one glycoprotein. Currently, only seven antigens have been defined . Also, neutrophil autoantibodies frequently react with some HNAs, especially to HNA-1 glycoproteins (CD16b, FcγRIIIb) .
This article is a review on neutrophil alloantigen in the Taiwanese population, and case investigations related to allo/auto neutrophil antibodies over a period of 20 years in the immunohaematology reference laboratory of the Mackay Memorial Hospital, Taiwan. A few case studies in this review were obtained from other hospitals in Taiwan who had requested granulocyte antibody testing.
HNA alloantigen distribution in Taiwan
Alloantigen distribution usually varies greatly between populations. In 1994, Dr. Marie Lin first published a study on neutrophil-specific antigen frequencies among Taiwanese using the granulocyte immunofluorescence test (GIFT) . HNA-1a antigen (92%), formerly called NA1, was more frequent than HNA-1b (formerly NA2) (52%). The distribution of HNA-1 among Taiwanese is similar to that seen in Japanese, but is very different to the distribution found in Caucasian or African where HNA-1b is usually more prominent than HNA-1a. In Taiwan, the frequency of HNA-2a (formerly NB1) reached 99%.
In year 2000, a sequence-specific primer method based on polymerase chain reaction (PCR-SSP) was used to genotype HNA-1 in Taiwanese populations . The allele frequency of FCGR3B*1 (allele encoding HNA-1a) was twice that of FCGR3B*2 (allele encoding HNA-1b) among the Taiwanese or the aboriginal population. This distribution was in concordance with the gene frequencies estimated from the 1994s serological typing results (0·68 for HN A-1a and 0·31 for HNA-1b). In addition, FCGR3B*3 (allele encoding HNA-1c, formerly SH) which is found in Caucasian and African groups, was not detected among Taiwanese, whereas the HNA-1 null phenotype was seen in one individual among 326 Taiwanese (<0·3%) and in three individuals among 98 Ami aboriginal people (∼3%).
FCGR3B gene duplication has been reported in Europeans with HNA-1 (a+b+c+) phenotype . HNA-1 null individuals (homozygous FCGR3B gene deletion) have also been seen in other populations worldwide. It was hypothesized that FCGR3B gene duplication or deletion is due to an unequal crossing-over between highly homologous sequence regions and consequently were assumed to be a rare event. This assumption appeared to be correct as for a long time gene duplication or deletion was hard to detect using traditional PCR method. More recently, we used quantitative PCR method and multiplex ligation-dependent probe amplification (MLPA) method to determinate the FCGR3B gene copy numbers. Surprisingly, our preliminary results showed that individual with FCGR3B gene copy number of one (deletion) was very common (14·9%). Accordingly, the HNA-1 null individual homozygous for FCGR3B gene deletion can be estimated to be approximately 1 in 180 (0·55%) which is not as rare as previously expected. On the other hand, individuals with FCGR3B gene copy number of three (duplication) were also frequently seen (14·5%). These individuals were initially genotyped as either homozygous FCGR3B*1, homozygous FCGR3B*2, or heterozygous with FCGR3B*1 and FCGR3B*2. This genotype pattern assumes two possible gene duplications: either two FCGR3B*1 alleles or two FCGR3B*2 alleles, and indicates that unequal crossing-over (deletion or duplication) is a frequent random event.
Alloimmune neonatal neutropenia
Granulocyte-specific alloantibodies were first reported from a family with multiple cases of neonatal neutropenia by Lalezari et al. . The prevalence of HNA-1 alloantibodies formation during pregnancy was estimated to be 0·6%  or 1·1%  in Caucasians, where the incidence of ANN due to the HNA-1 alloantibodies was much less frequent (0·016%  or <0·1% ).
HNA-1 alloantibodies induction rate during pregnancy was not systematically investigated in Taiwan. In a search among our previous laboratory requests of neonatal cases for granulocyte antibodies testing over a 10 year period (1997–2006), only four cases showed neutrophil alloantibodies in the mother and baby’s sera. In case number one, a mother of twins had a HNA-1 null phenotype, and she had isoantibodies against FcγRIIIb receptor. Case number two was due to maternal HNA-1b antibodies. Cases number three and four are associated to HNA-1a alloantibodies with babies presenting omphalitis.
Mothers homozygous for either HNA-1a (47·7%) or HNA-1b (10·1%) are efficient in inducing alloantibodies when the baby has counterpart alloantigens. However, the number of neonatal neutropenia caused by HNA-1 alloantibodies has rarely been seen in our clinical practice. It is possible that ANN was not always effectively diagnosed as the affected babies had no clear-cut sign of infections; alternatively, the immunogenicity of HNA-1a/-1b antigens is not strong in Taiwanese. There was far less mothers with HNA-1 null phenotype (0·55%) than mothers homozygous HNA-1a or -1b. However, one of the four ANN cases was a delivery from an FcγRIIIb deficient mother. Although we had a small number of cases, it appears that FcγRIIIb deficient mothers are more immunogenic against HNA-1 than mothers homozygous for HNA-1a or -1b.
Primary autoimmune neutropenia in children
In our 2009 publication , 55 children diagnosed with primary autoimmune neutropenia (PAIN) in a 13 years’ period (1993–2006) were retrospectively investigated. The study looked at the clinical course, autoantibody specificity, and a human leucocyte antigen (HLA) association study.
Among 55 patients, with mean age of onset of 9·8 months (range 4–28 months), the mean absolute neutrophil count at time of first diagnosis was 190 per μl. We had only 24 patients whose records were complete, the mean age of resolution was 22·5 months (range 13–44 months) and all patients recovered spontaneously. Recorded infections were usually mild affecting skin and upper respiratory tract.
Study on autoantibody specificity, showed that, anti-HNA-1a was commonly seen in our 41 patients (74%), and only one child (2%) had anti-HNA-1b. Target antigens could not be determined on 13 patients (24%) as their sera also reacted with all HNA-1a and -1b positive neutrophils. We arranged to test four of these sera for HNA-1 null neutrophils. All gave negative results. This suggests that the target of the autoantibodies was the FcγRIIIb receptor (HNA-1 glycoprotein).
To further understand the aetiology of PAIN, 31 cases were genotyped for HLA-DRB1 and -DQB1 genes. HLA-DQB1*05:03 was found to be significantly increased in PAIN children (0·177) compared to the control group (0·034) (odds ratio 6·48; P = 0·0002; Pc = 0·003). A previous German report  showed autoimmune neutropenia in children was associated with HLA-DR2 (later split as DR15 and DR16) and possibly also to DR6 (later split as DR13 and DR14). In our dataset, there appeared to be a weak association with DRB1*16:02 (OR 2·7; P = 0·0318) and HLA DRB1*14:05 (OR 5·89; P = 0·0099) but when corrected for the total number of alleles seen in the study (31 alleles), the corrected values for P (Pc) were not significant. A larger case–control study would be needed to confirm this weak association. HLA-DQB1 loci belong to the HLA class II gene system; as such their encoded proteins are expressed on the membrane of the antigen presenting cells, and are recognized by T cell receptors on T helper cells to modulate the immune response. It is most likely that DQB1*05:03 has an important role in the development of autoimmune antibody. Nonetheless, the molecular mechanism associated DQB1*05:03 and HNA-1 autoantibody production remains poorly understood and should be investigated further to elucidate the underlying cause of the pathogenesis of PAIN.
TRALI case investigation
Non-haemolytic transfusion reactions (NHTRs) are the most common transfusion reactions. They include febrile reactions, allergic reactions, bacterial sepsis and pulmonary transfusion reactions. TRALI is rare, but is the most serious NHTRs as it frequently causes mortality and morbidity after blood transfusion. Neutrophil alloantibodies are considered to be a causing factor of TRALI. A survey in our hospital (December 1999 to May 2000) showed a transfusion reaction incidence of 2·7% (358 cases in 12 822 patients) with most patients developing febrile reactions and urticaria. Only three cases among the 358 (<1%) developed respiratory distress . Few TRALI cases have been investigated in the past 20 years in our reference laboratory. The GIFT method was routinely used to screen for granulocyte antibodies, while lymphocytotoxicity test and solid-phase red cell adherence test were used to detect HLA and/or platelet antibodies. Although antibodies from donors or recipients against granulocytes, platelets and lymphocytes were rarely detected, two cases indicated possible antibody-mediated TRALI after packed red blood cells transfusion. The first case was due to a donor serum with antibodies against all our panels of granulocytes. Case two had no detectable specific antibodies against any cells; however, serum immunoglobulin level in the patient was dramatically decreased after transfusion reaction . It is not understood at this stage which antigen system or part of the immune system causes the immunoglobulin level to fall. No cognate antigen has yet been determined in these two cases. Further study is needed.
No potential conflict of interests to declare.